nabakin

8 ms·

61.
by nabakin 2y ago
Honestly, I don't understand why the threshold for war seems to be ground invasion. If you're shooting missiles at another country, that's war too. Israel is at war with Syria, Iran, and Lebanon imo. Not to mention Hamas.
62.
by nabakin 2y ago
You might be confusing it with Lebanon, but Syria has been bombed by Israel and is pretty unstable in general so it's impressive they were able to do this research regardless. People shouldn't be downvoting you for asking a questi
63.
by nabakin 2y ago
I think the nuance here is while we should continue to push for WFH, we should also be aware of the feelings of people around us when we speak from our privileged position. This doesn't just apply to WFH, we should keep other people&#x
64.
by nabakin 2y ago
Both represent the working class, just different subsets. Rural working class vs urban working class.
65.
by nabakin 2y ago
Why would that be necessary to prove the series has top performance?
66.
by nabakin 2y ago
I stand corrected, that's incredible. https://www.cpubenchmark.net/singleThread.html https://www.cpu-monkey.com/en/cpu_benchmark-cinebench_r23_si... https://www.cpubenchmark.net/hi
67.
by nabakin 2y ago
In power efficiency maybe, but not top performance
68.
by nabakin 2y ago
It does and it gets even worse when you realize those stats are only true under very specific circumstances, not typical computer usage. If you benchmarked based on typical computer usage, I think you'd only see gains of 5% or less.
69.
by nabakin 2y ago
In that case, wouldn't it have to be more complex than just a particle which creates gravity? Something which is compatible with the finding of spacetime curvature?
70.
by nabakin 2y ago
> I read/heard a great idea recently, what if gravity is an emergent quantity like heat? Maybe dense fermions just radiate gravitons just like a hot mass radiates photons? Gravity is the result of spacetime curvature, if I remember
71.
by nabakin 2y ago
It's interesting if the spike only happened on CrowdStrike computers but I'm not sure if this article has checked traffic logs on non-CrowdStrike computers and confirmed there was no spike. Even if they did, I agree there's l
72.
by nabakin 2y ago
Thanks, that's some really bad accuracy/performance
73.
by nabakin 2y ago
Are you sure the chat history is being passed when the second message is sent? That looks like the kind of response you'd expect if it only received the prompt "in python" with no chat history at all.
74.
by nabakin 2y ago
> I think Stable Diffusion was first to release a SOTA model (August 2022) that worked locally, not in language but image generation, but it set the tone for Meta. LLaMA only came in February 2023. Sure, I was only covering LLMs though.
75.
by nabakin 2y ago
I've been following the state of things, but I'm not sure which ones are the best. There's Meta's CodeLlama[1], Mistral's Codestral[2], DeepSeek AI's DeepSeek-Coder-V2-Instruct[3], CodeGemma[4], Alibaba's
76.
by nabakin 2y ago
Here's a summary of what's happened the past couple of years and what tools are out there. After ChatGPT released, there was a lot of hype in the space but open source was far behind. Iirc the best open foundation LLM that existe
77.
by nabakin 2y ago
Only thing I would change is the shuffling of characters is different for each user which can give a user an advantage over another depending on how they happen shuffle for them
78.
by nabakin 2y ago
> there's no CodeGPT, its just GPT4 Codex[1] is OpenAI's CodeGPT. It's what powers GitHub Copilot and it is very good but not publicly accessible. Maybe they don't want something else to outcompete Copilot. [1] https
79.
by nabakin 2y ago
Not sure about the low latency aspect, but I've seen everything else you mentioned with SSML. Also, I can't find where Greg said that, could you point me to it?
80.
by nabakin 2y ago
Azure Speech tts is capable of speeding up, slowing down, sarcasm, etc with SSML. I wouldn't be surprised if it's what OpenAI is using on the backend.
81.
by nabakin 2y ago
Azure Speech tts is capable of doing this with SSML. I wouldn't be surprised if it's what OpenAI is using on the backend.
82.
by nabakin 2y ago
Seems about as good as Azure's Speech Service. I wonder if that's what they are using behind the scenes
83.
by nabakin 2y ago
Exactly. Falcon-180b had a lot of hype at first but the community soon realized it was nearly worthless. Easily outperformed by smaller LLMs in the general case. Now they are back and claiming their falcon-11b LLM outperforms Llama 3 8b. I
84.
by nabakin 2y ago
> An alarming number of people think their chats were leaked Easily explained by direct access to the phone or Pegasus (or Pegasus-like) spyware. Both of which Telegram is also vulnerable to.
85.
by nabakin 2y ago
Or some sort of spyware like Pegasus
86.
by nabakin 2y ago
Maybe for battery life, but definitely not when it comes to CPU/GPU performance. Tbf, no chip company is, but Apple is particularly egregious. Their charts assume best case multi-core performance when users rarely ever use all cores at
87.
by nabakin 2y ago
GPT-4 has been updated multiple times since then. Anthropic has been trading blows with those updates to GPT-4.
88.
by nabakin 2y ago
Not ahead of Anthropic. Anthropic is trading blows with the best GPT-4 models these days. For the open weight/source community I'd put it at 3-6 months behind and catching up. Llama 3 405B might catch the community up unless OpenA
89.
by nabakin 3y ago
ffmpeg wasm is very slow in the browser. Users would still use those terrible websites because they wouldn't be terribly slow
90.
by nabakin 3y ago
Not only has Microsoft put a button that opens Edge in the taskbar, but that button opens on hover . Their anticompetitive practices are getting more ridiculous by the day. When I tell you I want to use Firefox, stop asking me if I want to

More ›